The proposed research involves the following subject areas: survival analysis, classification and non-linear regression, and statistical deconvolution and regularization. Among the topics we plan to consider are an investigation of the efficiencies of goodness-of-fit procedures with censored data, a modification of the Gehan test or the Mantel-Haenszel test for the sequential analysis of randomly censored data, an investigation of various statistical procedures for survival analysis derived from Cox's partial likelihood, a continuation of research in multivariate estimation problems especially concerning multivariate regression functions and error rates in discriminant analysis, a continuation of research into the problems of variable selection in estimation of error rates with recursive partitioning schemes, a continuation of a theoretical and empirical investigation of methods for deconvolving estimated probability density functions. For example, Professors Koziol and Petkau this past year completed a numerical study of various sequential tests for the equality of survival distributions with singly censored data.